Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McDannold, N. J.
Right arrow Articles by Hynynen, K. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McDannold, N. J.
Right arrow Articles by Hynynen, K. H.
(Radiology. 1999;211:419-426.)
© RSNA, 1999

Experimental Studies

Determination of the Optimal Delay between Sonications during Focused Ultrasound Surgery in Rabbits by Using MR Imaging to Monitor Thermal Buildup in Vivo1

Nathan J. McDannold, BS, Ferenc A. Jolesz, MD and Kullervo H. Hynynen, PhD

1 From the Department of Radiology, Division of Magnetic Resonance Imaging, Brigham and Women's Hospital, Harvard Medical School, LMRC, 007c, 221 Longwood Ave, Boston, MA 02115 (N.J.M., F.A.J., K.H.H.), and the Department of Physics and Astronomy, Tufts University, Medford, Mass (N.J.M.). Received January 20, 1998; revision requested March 17; final revision received August 7; accepted October 6. Supported by National Cancer Institute grants RO1:CA46627 and PO1:CA67165 and a grant from GE Medical Systems. Address reprint requests to N.J.M.


   Abstract
TOP
 Abstract
Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
PURPOSE: To use magnetic resonance (MR) imaging to monitor thermal buildup and its effects in treated tissues during sequentially delivered sonications in vivo to optimize the intersonication delay for any set of ultrasound and tissue parameters.

MATERIALS AND METHODS: Sequential sonications were delivered next to each other in both thighs in 10 male New Zealand white rabbits. The time between sonications was 11–60 seconds. Phase-difference MR imaging was used to monitor temperature rise, which was used to estimate the thermal dose delivered to the tissue. T2-weighted and contrast agent–enhanced T1-weighted imaging were used to gauge the extent of tissue coagulation.

RESULTS: With a short intersonication delay (11–40 seconds), the estimated temperature rise and the extent of tissue coagulation increased dramatically in subsequent sonications. However, when the delay was long (50–60 seconds), the size and shape of the destroyed tissue with subsequent sonications was uniform, and the temperature buildup was substantially lower.

CONCLUSION: MR imaging can be used to monitor thermal buildup and its effects due to sequential, neighboring sonications in vivo to produce evenly shaped regions of tissue coagulation. The temperature information obtained from the monitoring can be used to optimize the intersonication delay for any set of ultrasound and tissue parameters.

Index terms: Magnetic resonance (MR), artifact, 44.93 • Magnetic resonance (MR), phase imaging, 44.121411, 44.121412 • Magnetic resonance (MR), thermometry, 44.1214 • Ultrasound (US), focused, 44.1298 • Ultrasound (US), experimental, 44.1298 • Ultrasound (US), therapeutic, 44.1298


   Introduction
TOP
 Abstract
 Introduction
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
Focused ultrasound has been shown to be a promising noninvasive surgical method for the destruction of deep-seated tissue (1). Previous studies (2,3) have shown that high-intensity, short-duration sonications provide nearly perfusion-independent temperature rise and lesion formation. However, when closely spaced sonications are delivered without a sufficient intersonication delay, thermal buildup results because of cumulative ultrasound exposure near the transducer (4–7). While the temperature rise near the focus is not sufficient to cause tissue destruction with a single sonication, when sonications are delivered sequentially close to one another, the additive temperature rise can cause increasingly larger lesions.

Furthermore, with sufficient thermal buildup, bubbles can form owing to the boiling of water in the tissue. These gas bubbles can cause scattering and reflection of the ultrasound beam. Scattering and reflection can cause the deposition of the ultrasound energy in front of the bubbles, thereby moving the region of greatest temperature rise nearer still to the transducer (8). Because of this effect, after several sonications the coagulated tissue volumes have different shapes and sizes than those produced with the first sonications. Thus, patient safety demands that focused ultrasound delivery to a large tissue volume have a sufficient intersonication delay to minimize thermal buildup. Currently, this delay is selected on the basis of worst-case estimates that are based on findings of studies (4–7) in tissue with a low rate of perfusion, which results in many cases in unnecessarily long delays.

Thermal buildup is strongly dependent on the perfusion rates in different tissues as well as on ultrasound parameters (geometry, focal volume, sonication time, power) (4). Previous studies (4–7) have demonstrated thermal buildup effects and have suggested guidelines for intersonication delays for different tissue and ultrasound parameters. The results from such studies may not be optimal for a particular treatment because they represent worst-case scenarios. A shorter delay may be possible due to differences in perfusion. Furthermore, the studies (4–7) on thermal buildup in which image guidance was not used depended on simulations, invasive temperature measurements at discrete points, and posttreatment dissection.

To minimize the intersonication delay and thus the whole treatment time, the temperature buildup should be measured online during each treatment. This temperature information can then be used to control the required sonication delay online. Magnetic resonance (MR) imaging has been shown to be useful in monitoring and guiding focused ultrasound treatments (9,10). MR imaging can be used to plan the treatment, localize the ultrasound focus, characterize the temperature rise during the therapy, and provide follow-up analysis of the relative success of the treatment (11).

Short-duration sonications are desirable when used in conjunction with MR imaging thermometry, because imaging can occur during the entire sonication with less concern about magnetic field drift. Thus, the use of multiple short-duration, high-intensity sonications is a desirable approach to the focused ultrasound surgery of large tissue volumes because one can coagulate consistently defined regions and can accurately monitor the temperature rise with MR imaging thermometry. Findings of a recent study (12) in which imaging methods and ultrasound parameters similar to those described in this article were used to evaluate the focused ultrasound treatment of tumors implanted in rabbit thigh muscle showed that MR imaging did not depict excessive thermal buildup after several sequential sonications that were separated by a single 50-second intersonication delay.

In this study, the use of MR imaging to find an optimal intersonication delay by monitoring the thermal buildup in vivo was investigated in depth. MR imaging thermometry was used to characterize the temperature rise over the range of the entire ultrasound field during the delivery of sequential neighboring, high-intensity, short-duration sonications in rabbit thigh muscle. Several intersonication delays were investigated.

We examined whether imaging could be performed over the entirety of the sonications and cooling and whether any artifacts induced by the motion of the transducer or patient could be corrected. The temperature profiles were used to estimate the thermal dose delivered. Temperature measurements, dose estimates, and postsonication imaging were used to quantify the thermal buildup effects. These temperature measurements can be used to optimize the intersonication delay for any given treatment. Unlike in previous studies (4–7) on thermal buildup in which MR imaging monitoring was not used, such a method provides online information about thermal buildup effects over the range of the entire ultrasound field for any set of ultrasound and tissue parameters regardless of the perfusion rate. The purpose of this study was to use MR imaging to monitor thermal buildup and its effects in treated tissues during sequentially delivered sonications in vivo to optimize the intersonication delay for any set of ultrasound and tissue parameters.


   MATERIALS AND METHODS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 References
 
Experimental Apparatus
The ultrasound fields were generated by a spherically curved air-backed, single-focus transducer (GE Medical Systems, Milwaukee, Wis) that had a diameter of 100 mm, a radius of curvature of 80 mm, and a resonant frequency of 1.5 MHz. Acoustic power measurements and the half-intensity beam length and diameter were determined by using a method described by Hynynen et al (9). The half-intensity beam length and diameter were 4.8 and 1.0 mm, respectively. The transducer was mounted in a computer-controlled, mechanically driven, MR imaging–compatible positioning system (GE Medical Systems); an earlier version of the positioning system is described by Cline et al (13).

Focused ultrasound was delivered to both thighs in 10 male New Zealand white rabbits. Before treatment, the hair on the thighs was removed with an electric shaver and hair remover. The animals were anesthetized with a mixture of sodium xylazine hydrochloride (12 mg per kilogram of body weight per hour; Fermenta Animal Health, Kansas City, Mo) and ketamine hydrochloride (48 mg per kilogram of body weight per hour; Fort Dodge Laboratories, Fort Dodge, Iowa). The anesthetic was administered as an intramuscular injection every hour. The animal body temperature was constantly monitored with a copper-constantan rectal thermocouple (constructed in-house). The rabbits were provided housing, food, and veterinary care according to National Institutes of Health and Harvard Medical School Guidelines (14).

The rabbits were placed on a Plexiglas tray above the positioning system, and the thigh sat in a hole above the transducer that contained a bag of deionized, degassed water. The bag rested on a polyvinyl chloride membrane above the transducer, which was also submerged in degassed water. To provide proper coupling, a layer of degassed water was poured between the membrane and the plastic bag. A 12.7-cm-diameter, receive-only surface coil was attached beneath the Plexiglas tray below the thigh to improve the signal-to-noise ratio. The whole apparatus was then placed in a 1.5-T clinical MR imager (Signa; GE Medical Systems). A diagram of the experimental setup is shown in Figure 1.



View larger version (79K):
[in this window]
[in a new window]
  		Figure 1. Diagram shows the experimental setup for the delivery of focused ultrasound in rabbit thigh muscle in vivo. PVC = polyvinyl chloride, RF = radio frequency.

 
Imaging
The focused ultrasound targets were planned with a fast spin-echo T2-weighted imaging sequence: 2,000/95 (repetition time msec/echo time msec); echo train length, eight; field of view, 16 cm; section thickness, 3 mm; matrix size, 256 x 256; number of signals acquired, two; imaging time, 264 seconds. Phase-difference imaging was used to localize the focus and measure temperature rise by proton-resonant frequency shifts. The phase shift was calculated by using a fast spoiled gradient-echo sequence: 26.6/16.8; flip angle, 30°; bandwidth, 4 kHz; field of view, 16 cm; matrix size, 256 x 128; section thickness, 3 mm; imaging time per image, 4.5 seconds (15,16).

The temperature dependence of the proton-resonant frequency shift in rabbit skeletal muscle is 0.00909 ppm/°C and is linear above the tissue coagulation threshold temperature (17). With an echo time of 16.8 msec, the temperature dependence of the spoiled gradient-echo sequence was 0.061 radians/°C (16). The frequency shift was calculated by multiplying the phase shift by 2{pi}/(echo time).

The optimal echo time for phase-difference imaging in muscle occurs when the echo time is equal to T2* (25 msec in muscle) (16). The use of a shorter echo time lessens the repetition and imaging times, providing better temporal resolution and helping to avoid phase wraparound. A further reduction in phase wraparound was achieved by using a complex phase-subtraction method (16). Phase wraparound is caused by phase changes larger than {pi} radians, corresponding to a change in temperature of 51.3°C for an echo time of 16.8 msec. For phase changes greater than {pi}, a phase-unwrapping scheme was used and is described in the Appendix.

A large temperature rise can also result in cavitation because of the boiling of water in the tissue. When the high-intensity ultrasound focus is at a tissue interface, cavitation can also result (18). When cavitation occurs, the phase change no longer represents a temperature rise because of a randomization of the phase development (19). The measurements in regions where cavitation occurred were ignored.

After treatment, T2-weighted and T1-weighted images (before and after contrast enhancement) were used to gauge the extent of the treatment. The parameters used for the T1-weighted imaging were 500/17; echo train length, four; field of view, 16 cm; section thickness, 3 mm; matrix size, 256 x 256; number of signals acquired, three; imaging time, 97 seconds. The contrast agent used was gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) at a concentration of 0.3 mmol per kilogram of body weight. After the bolus of contrast agent was injected in the ear veins of the rabbits, images were obtained until the enhancement reached its maximum. Images obtained after the injection of the contrast agent were subtracted from images obtained before the injection of the contrast agent. Contrast agent–enhanced images are good indicators of tissue necrosis (20).

Focused Ultrasound Delivery and Image Analysis
Sequential sonications were delivered in both thighs of 10 rabbits. The sonications were at 30-W acoustic power for 10 seconds, with the focus 20–30 mm beneath the skin. The intersonication time was 11–60 seconds. The largest possible spacing that would provide overlapping lesions in a volume treatment for an estimated 4-mm lesion was used (4/{surd}2 = 2.8-mm spacing).

In 18 thighs, one to three rows of sonications were delivered, as shown in Figure 2. Twenty-seven rows were analyzed. In each row, seven to 10 sonications were delivered, with an intersonication delay of 11–60 seconds (from row to row). The separation between rows was chosen so that one row did not overlap with the next.



View larger version (51K):
[in this window]
[in a new window]
  		Figure 2. Diagram shows the method of sonication. Sequential neighboring sonications were delivered to the rabbit muscle in rows. The intersonication delay was varied from row to row. The ultrasound fields overlap, and thermal buildup can result.

 
Phase-difference images were acquired before, during, and after all sonications. For each row of sonications, 70–150 phase subtractions were performed, depending on the intersonication delay and number of sonications, yielding 7.5–22.5 minutes of continuous imaging. The sonications were imaged parallel to the ultrasound beam in the 27 rows. The spherically curved transducer produces a heating pattern in the shape of a cone, and the direction of the ultrasound beam (Fig 2) is defined as the direction of this cone (ie, perpendicular to the face of the transducer). Images were acquired in a plane parallel to the direction of the ultrasound beam because the thermal buildup progressed toward the transducer and was symmetric about the axis of the ultrasound beam. Imaging parallel to the direction of the ultrasound field yields a temperature map of a plane that extends over the range of the entire ultrasound field.

Images also were acquired during the motion of the positioning system (approximately 10 seconds). This motion produced an artifact in the images because of the metal contained in the positioning system, but this artifact was corrected (see Appendix). Because of the relatively long time during which the images were obtained, magnetic field drift became a concern. Measurements obtained with the MR imaging system show a field drift corresponding to 0.17 radians/h ± 0.013 (SD; 2.8°C/h ± 0.21 with an echo time of 16.8 msec) (21). However, even with the longest imaging time used in this study, the drift would be only about 1°C, which is less than the noise in the images.

The temperature elevation history and the rabbits' body temperature were used to estimate the thermal dose delivered to the tissue during each series of sonications (22). A thermal dose equivalent to 43°C for 240 minutes was used to determine tissue destruction (17). A sublethal thermal dose was also used to observe the increase in low-level thermal dose as a function of intersonication delay. The thermal dose estimates were compared to the postsonication images. All data analysis was performed after the completion of the experiments.


   RESULTS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 APPENDIX
 References
 
Figure 3a shows a T2-weighted image of the lesions formed by eight successive sonications, with an intersonication delay of 50 seconds. The sonications proceeded from left to right in the Figure, and the spacing between sonications was 2.8 mm. Figure 3b shows a phase-difference image obtained at the peak temperature rise of the first sonication in the same thigh. The temperature rise at the focus is clearly shown. Figure 3c shows the temperature elevation during the last sonication in the row after motion artifact correction was performed as described in the Appendix. The images in Figure 3 were acquired parallel to the ultrasound beam.



View larger version (85K):
[in this window]
[in a new window]
  		Figure 3. Focused ultrasound delivered in rabbit thigh muscle in vivo. A, T2-weighted (2,000/95) MR image obtained after seven sonications were delivered with an intersonication delay of 50 seconds. B, Temperature-sensitive phase-difference spoiled gradient-echo MR image (26.6/16.8) of the first sonication delivered to the tissue at the peak temperature rise. C, Temperature-sensitive phase-difference spoiled gradient-echo MR image (26.6/16.8) of the last sonication delivered to the tissue. Some thermal buildup between B and C is visible, but it is small.

 
Figure 4 shows the temperature rise as a function of time for 10 sequential sonications delivered in rows, with two different intersonication delays. Each line in the plots follows the temperature development of a small volume beneath the ultrasound focus for each of the 10 sonications to measure the thermal buildup in the near field. A distance of 5 mm was chosen because it is on the order of the half width of the length of the focus of the transducer.



View larger version (41K):
[in this window]
[in a new window]
  		Figure 4. Graphs show temperature estimates as a function of time. Top: Graph for an intersonication delay of 11 seconds. Bottom: Graph for an intersonication delay of 50 seconds. Each line represents the temperature development of a region of tissue 5 mm below the focal plane during 10 sonications.

 
For example, the line in each graph in Figure 4 that shows the first peak in temperature rise indicates the temperature development of a volume below the first sonication in the row. The magnitude of the peak temperature rise of the sonications with an 11-second intersonication delay increases relatively rapidly compared with that of a 50-second intersonication delay. Each line represents an mean of an approximately 27-mm3 region of tissue 5 mm below the ultrasound focal plane (mean of 5 x 5 voxels; voxel size, 0.625 x 0.625 x 3 mm). The SDs from the averaging are not shown. The mean deviation was 1.5°C ± 0.6. The mean fluctuation per voxel in the images away from the heating was ±0.15 radians (corresponding to ±2.5°C with an echo time of 16.8 msec).

The mean estimated temperature rise at 13.5 seconds, which was the duration of three image acquisitions, after one, three, five, and seven sonications is plotted for a distance of 5 mm below the focal plane in Figure 5. In this Figure, the temperature rise is shown across a region of tissue (each line is a mean of three rows of voxels) perpendicular to the ultrasound beam. The increase in peak temperature rise and the spread of thermal buildup over space is evident in the Figure. With an intersonication delay of 11 seconds, the peak temperature rise after seven sonications was 156% greater than the peak temperature of the first sonication. In contrast, with an intersonication delay of 50 seconds the peak temperature was 51% greater than the peak temperature of the first sonication and remained constant to within 5% after the third sonication. This temperature increase is also seen in Figure 4.



View larger version (33K):
[in this window]
[in a new window]
  		Figure 5. Graphs show spatial temperature profiles 5 mm below the focal plane 13.5 seconds after one, three, five, and seven sonications for an intersonication delay of 11 seconds (top) and 50 seconds (bottom).

 
Figure 6 shows a summary of the temperature rise as a function of intersonication delay at three distances below the focal plane. These temperatures were measured between sonications (13.5 seconds after seven sonications). The thermal buildup dramatically increases as the intersonication delay is decreased. The temperature rise in regions where there was cavitation was not used in any of the image analyses.



View larger version (21K):
[in this window]
[in a new window]
  		Figure 6. Graph shows estimates of temperature rise as a function of intersonication delay for three depths below the focal plane.

 
The measurements of temperature rise over time and space were used to estimate the thermal dose delivered to the tissue. Figure 7 shows contours indicating estimated thermal doses equivalent to 10 and 240 minutes at 43°C superimposed on postsonication T2-weighted and contrast-enhanced T1-weighted subtraction images with two different intersonication delays. On the contrast-enhanced images, regions of tissue coagulation appear black because there is no uptake of contrast agent. On the T2-weighted images, the edema is bright, and regions of tissue destruction are dark. The regions of tissue destruction were more visible on the contrast-enhanced images. The contours match the regions of tissue coagulation fairly well, given that the tissue can swell and move because of the heating and tissue damage induced by the focused ultrasound. The images in Figure 7 were acquired in the same plane as were the phase-subtraction images (parallel to the ultrasound beam), which corresponds to the center of the focused ultrasound-induced lesions.



View larger version (138K):
[in this window]
[in a new window]
  		Figure 7. Thermal dose estimates superimposed on postsonication T2-weighted (2,000/95; top) and contrast-enhanced T1-weighted (500/17; bottom) MR images for two intersonication delays. Left side: Images obtained after an intersonication delay of 11 seconds. Right side: Images obtained after an intersonication delay of 60 seconds. The contrast-enhanced images are subtractions of images acquired before and after the bolus of contrast agent was injected. The thermal dose contours indicate tissue coagulation (240 equivalent minutes at 43°C, thick solid line) and a sublethal dose (10 equivalent minutes at 43°C, thin dotted line). The images were acquired parallel to the ultrasound beam.

 
Figure 8 shows T2-weighted images obtained perpendicular to the ultrasound beam for two different intersonication delays. The images were acquired through the focal plane. Again, the extent of tissue destruction increased with a shorter intersonication delay. The images in Figures 7 and 8 should correspond to the center of the lesions when evenly shaped.



View larger version (157K):
[in this window]
[in a new window]
  		Figure 8. Postsonication T2-weighted (2,000/95) MR images acquired perpendicular to the ultrasound beam for two intersonication delays. The size of each subsequent lesion grew owing to the thermal buildup with the 11-second delay, while with the 50-second delay, the lesions were evenly shaped.

 

   DISCUSSION
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
APPENDIX
 References
 
Postsonication imaging showed that evenly shaped and evenly sized lesions could be created with sequential, neighboring, high-intensity sonications when the intersonication delay was sufficient to minimize the effects of thermal buildup. Phase-difference imaging was appropriate for quantification of the temperature buildup between sonications and the thermal dose delivered during the sonications. This information can be used online to optimize the delivery of focused ultrasound for any given set of ultrasound and tissue parameters. With the tissue and ultrasound parameters used in this study, a minimal delay of 50–60 seconds was necessary to create evenly sized lesions. Muscle has a relatively low perfusion rate, so the results from this study represent the worst-case scenario.

The proper intersonication delay is difficult to calculate for a given set of ultrasound parameters. The delay depends not only on the ultrasound parameters (geometry, focal volume, applied power, sonication time) but also on perfusion rates in the tissue (4). Previous studies (4,6) have shown a wide range of intersonication delays depending on these parameters. Fan and Hynynen (5) have shown in simulations with the same transducer geometry and sonication time as those used in this study, but with an applied power of 20 W, that an intersonication delay of 40 seconds is necessary to avoid thermal buildup in tissue with a low perfusion rate. This result is close to the finding in the current study.

The use of MR imaging to monitor thermal buildup was demonstrated recently by McDannold et al (12) in an article describing the MR imaging–guided focused ultrasound treatment of tumors implanted in rabbit thigh muscle. The authors of that article (12) found that the 50-second intersonication delay used in the sonications did not cause excessive thermal buildup. It showed that the temperature rise 5 mm from the skin (1.5–2.5 cm below the focal plane) was less than 1°C after four sequential sonications and that the lesion size increased radially (in the direction perpendicular to the ultrasound beam) by 2 mm ± 2 (12). In that study (12), however, the buildup of the thermal dose was not monitored parallel to the ultrasound beam, and the temperature information was obtained for only a small part of the cooling time and not during the motion of the ultrasound positioning system. The results from the current study demonstrate that a more in-depth analysis of the thermal buildup yields an optimal intersonication delay of 50–60 seconds in muscle tissue.

While the postsonication images provide evidence after the fact that evenly shaped regions of tissue destruction can be created, temperature estimates can be used to monitor online the effects of thermal buildup. Temperature-sensitive images indicate the effects of thermal buildup between sonications. With the information obtained from the images, one can determine the magnitude and extent of the temperature buildup to find the optimal intersonication delay.

Perhaps the best method for determining the effects of thermal buildup is to use the temperature estimates to calculate the equivalent thermal dose delivered to the tissue (3,17). An online thermal dose calculation allows for an estimate of where the tissue is being destroyed. The growth in dose from the cumulative thermal buildup can also be estimated. A thermal dose estimate based on the temperature estimates from phase-difference images can be used to indicate this growth.

Although the data analysis in this study was performed after the completion of the experiments, there was no analysis performed that could not have been done online. In a clinical setting, online MR imaging feedback may indicate that the ultrasound parameters need to be changed to optimize the treatment. However, for the purposes of this particular study, online analysis was not needed because the ultrasound parameters were not altered during the sonications.

A further useful result was found because of the correction scheme required for the artifact induced in the images owing to the motion of the positioning system (see Appendix). This correction scheme demonstrates that the use of many sequential phase-subtraction images is feasible for monitoring temperature rise and that a small amount of motion or changes in the magnetic field are correctable.

Practical applications: MR imaging thermometry allows for online, noninvasive monitoring of thermal buildup in vivo. It allows for feedback to the operator of the focused ultrasound unit so that the appropriate intersonication delay can be tailored to a given set of ultrasound and tissue parameters. One may envision a treatment of a large volume in which the intersonication delay is subsequently shortened until the temperature rise and growth of coagulation determined on the basis of thermal dose estimates reaches a predefined threshold below the threshold that is safe for the patient.

Conversely, one might use a short intersonication delay and use the MR imaging feedback to determine the optimal locations for subsequent sonications as the regions of tissue destruction increase in size. With such an approach, the thermal buildup is used to the advantage of the operator of the focused ultrasound unit. The increasing size of the lesions generated would be taken into account in the choice of the location and the power of subsequent sonications. This combination of the increasing lesion size and short intersonication time could greatly speed up the treatment time of a large tissue region. Both of these methods use the feedback obtained from the MR imaging monitoring of temperature to ensure that patient safety is not jeopardized.

Such methods could dramatically speed up the treatment, because a large number of sonications is necessary to treat a large volume. Findings of a previous study (12) showed that it took a mean of 59 sonications (with the same ultrasound parameters) to treat tumors with a volume of 5–10 cm3. The ability to create evenly shaped lesions is essential for the success of focused ultrasound therapy and for the safety of the patient when multiple sonications are used to destroy a large tissue volume. Being able to monitor the extent of tissue destruction with MR imaging allows for the determination of a proper intersonication delay and an online method to verify that the lesions are evenly shaped and evenly sized.


   APPENDIX
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
References
 
Correction of Motion Artifacts
Phase-difference imaging takes advantage of the temperature dependence of the proton-resonant frequency. Any shift in phase is found by subtracting the phase of each voxel acquired during and after heating from the phase in an image acquired before the heating. The frequency shift is calculated by multiplying this phase shift by 2{pi}/(echo time). If the magnetic field changes after the baseline image is acquired or the animal moves, the measurement of the phase change, and therefore the temperature estimate, is not valid.

Many components in the MR imaging–compatible mechanical positioning system used in this study were metallic. Thus, their position affected the overall magnetic field in the imaging plane. The changes in images obtained before and after the system moved did not reflect the true changes in temperature but rather artifacts due to the position of the metallic components of the positioning device.

A scheme to correct these artifacts was therefore introduced. Images were subtracted from a single baseline, and the phase development in a region of tissue without heating was plotted against time to see when the artifact occurred. The artifact represented itself as discontinuities in the phase development over time. The images obtained directly after the artifacts developed were used as new baseline images.

To take into account any heating in the tissue prior to the transducer motion, the values in each pixel before the motion were added to the values in the voxel after the motion after spatial averaging was performed. This averaging took the mean of each voxel with its three nearest neighbors to eliminate cumulative effects of noise in the image over time. Where the phase gradients were not steep (such as times between sonications or away from the focus), this averaging had little effect. Some images were excluded because of artifacts induced by the motion of the positioning system itself. These artifacts were seen as peaks in the phase development or delays in time of the artifact development at different locations in space.

Experiments were performed to test this correction. Rabbits were placed above the positioning device, and no focused ultrasound was delivered. The positioning system was moved as in the sonication experiments. Figure A1 shows the phase development over time before and after the artifact correction with seven movements of the positioning system. The gaps in the corrected phase plot represent the images used for the new baselines obtained after the motion and the images that were excluded because of the motion artifact. The artifact is dramatic, and the correction works well. The slow decay in the corrected artifact was owing to either the cooling of the rabbit in the magnet or drifts in the magnetic field. This experiment was repeated in three rabbits.



View larger version (28K):
[in this window]
[in a new window]
  		Figure a1. Graph shows correction of artifact induced by motion of the mechanical positioning system. Phase development over time is shown with no delivery of focused ultrasound but with seven movements of the positioning system before and after the correction. Error bars indicate SD.

 
For the experiments with heating of tissue, the same procedure was followed in regions of tissue where there was no heating. In these regions, the phase shift was 0 radians to an accuracy of ±0.12 radians (corresponding to a temperature shift of ±2°C for an echo time of 16.8 msec), which is less than the noise in the phase-subtraction images.

It is important to note that this artifact is a direct result of the positioning system and not of the method used in this experiment. A nonmetallic positioning system can be used without this artifact. In prior experiments with a hydraulic positioning system, this artifact was not observed when thermal buildup was studied in much the same way in sonications delivered to tumors implanted in rabbit thigh muscles (12).

Furthermore, this correction can also be used for small motion during sonications. A new baseline can be chosen after the motion, and the temperature rise before the motion can be added to the temperature estimates determined from the subtractions after the motion.

Phase Unwrapping
Phase subtractions are accurate only to modulo ±{pi}. A phase development out of this range results in wraparound and an incorrect temperature estimate. To correct this wraparound, any phase jump greater than {pi} was corrected by adding its 2{pi} complement. Implicit in this correction was the assumption that the temperature rise did not result in a phase change faster than {pi} radians in one image acquisition, or 11.4°C/sec for an echo time of 16.8 msec and an imaging time of 4.5 seconds. This method for correction is similar to a method proposed to correct phase wrapping in conjunction with phase-difference imaging with stereotactic devices that acquire an additional image between the baseline and the image to be subtracted (23).


   Footnotes
 
Author contributions: Guarantor of integrity of entire study, K.H.H.; study concepts and design, K.H.H.; definition of intellectual content, K.H.H.; literature research, N.J.M.; experimental studies, N.J.M.; data acquisition and analysis, N.J.M.; statistical analysis, N.J.M.; manuscript preparation, N.J.M.; manuscript editing and review, K.H.H., F.A.J.


   References
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 

  1. Hynynen K. Focused ultrasound surgery guided by MRI. Sci Med 1996; 3:62-71.
  2. Billard BE, Hynynen K, Roemer RB. Effects of physical parameters on high temperature ultrasound hyperthermia. Ultrasound Med Biol 1990; 16:409-420.[Medline]
  3. Damianou C, Hynynen K. The effect of various physical parameters on the size and shape of necrosed tissue volume during ultrasound surgery. J Acoust Soc Am 1994; 95:1641-1649.[Medline]
  4. Damianou C, Hynynen K. Near-field heating during pulsed high temperature ultrasound hyperthermia treatment. Ultrasound Med Biol 1993; 19:777-787.[Medline]
  5. Fan X, Hynynen K. Ultrasound surgery using multiple sonications: treatment time considerations. Ultrasound Med Biol 1996; 22:471-482.[Medline]
  6. Malcolm AL, ter Haar GR. Ablation of tissue volumes using high intensity focused ultrasound. Ultrasound Med Biol 1996; 22:659-669.[Medline]
  7. Wan H, VanBaren P, Ebbini ES, Cain CA. Ultrasound surgery: comparison of strategies using phased array systems. IEEE Trans Ultrasonics Ferroelectrics Frequency Control Soc 1996; 43:1085-1098.
  8. Hynynen K, Darkazanli A, Unger E, Schenck JF. MRI-guided noninvasive ultrasound surgery. Med Phys 1993; 20:107-115.[Medline]
  9. Hynynen K, Freund WR, Cline HE, et al. A clinical noninvasive MRI monitored ultrasound surgery method. RadioGraphics 1996; 16:185-195.[Abstract/Free Full Text]
  10. Cline HE, Schenck JF, Watkins RD, Hynynen K, Jolesz FA. Magnetic resonance guided thermal surgery. Magn Reson Med 1993; 31:628-636.
  11. Cline HE, Hynynen K, Hardy CJ, Watkins RD, Schenck JF, Jolesz FA. MR temperature mapping of focused ultrasound surgery. Magn Reson Med 1994; 30:98-106.
  12. McDannold N, Hynynen K, Wolf D, Wolf G, Jolesz F. MRI evaluation of thermal ablation of tumors with focused ultrasound. JMRI 1998; 8:91-100.
  13. Cline HE, Hynynen K, Watkins RD, et al. Focused US system for MR imaging–guided tumor ablation. Radiology 1995; 194:731-737.[Abstract/Free Full Text]
  14. . Guide for the care and use of laboratory animals. Bethesda, Md: U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health, 1985; NIH publication no. 85-23..
  15. Ishihara Y, Calderon A, Watanabe H, et al. A precise and fast temperature mapping using water proton chemical shift. Magn Reson Med 1995; 34:814-823.[Medline]
  16. Chung A, Hynynen K, Cline HE, Colucci V, Oshio K, Jolesz F. Optimization of spoiled gradient-echo phase imaging for in vivo localization of focused ultrasound beam. Magn Reson Med 1996; 36:745-752.[Medline]
  17. Chung A, Hynynen K, Jolesz FA. Thermal dosimetry of a focused ultrasound beam in vivo by MRI. Med Phys ; (in press).
  18. Hynynen K. The threshold for thermally significant cavitation in dog's thigh muscle in vivo. Ultrasound Med Biol 1991; 17:157-169.[Medline]
  19. Hynynen K, Chung AH, Jolesz FA. In vivo detection of ultrasound induced cavitation using MRI (abstr) In: Proceedings of the Fifth Meeting of the International Society for Magnetic Resonance in Medicine. Berkeley, Calif: International Society for Magnetic Resonance in Medicine, 1997; 525.
  20. Hynynen K, Darkazanli A, Damianou C, Unger E, Schenck JF. The usefulness of contrast agent and GRASS imaging sequence for MRI guided noninvasive ultrasound surgery. Invest Radiol 1994; 29:897-903.[Medline]
  21. Kuroda K, Suzuki Y, Ishihara Y, Okamoto K, Suzuki Y. Temperature mapping using water proton chemical shift obtained with 3D-MRSI: feasibility in vivo. Magn Reson Med 1996; 35:20-29.[Medline]
  22. Sapareto SA, Dewey WC. Thermal dose determination in cancer therapy. Int J Radiat Oncol Biol Phys 1984; 10:787-800.[Medline]
  23. Sumanaweera TS, Glover GH, Hemler PF, et al. MR geometric distortion correction for improved frame-based stereotaxic target localization accuracy. Magn Reson Med 1994; 34:106-113.



This article has been cited by other articles:


Home page
 Am. J. Roentgenol.Home page
W. Hundt, E. L. Yuh, M. D. Bednarski, and S. Guccione
Gene Expression Profiles, Histologic Analysis, and Imaging of Squamous Cell Carcinoma Model Treated with Focused Ultrasound Beams
Am. J. Roentgenol., September 1, 2007; 189(3): 726 - 736.
[Abstract] [Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
C. T.W. Moonen
Spatio-Temporal Control of Gene Expression and Cancer Treatment Using Magnetic Resonance Imaging-Guided Focused Ultrasound
Clin. Cancer Res., June 15, 2007; 13(12): 3482 - 3489.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
K. Hynynen, O. Pomeroy, D. N. Smith, P. E. Huber, N. J. McDannold, J. Kettenbach, J. Baum, S. Singer, and F. A. Jolesz
MR Imaging-guided Focused Ultrasound Surgery of Fibroadenomas in the Breast: A Feasibility Study
Radiology, April 1, 2001; 219(1): 176 - 185.
[Abstract] [Full Text]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McDannold, N. J.
Right arrow Articles by Hynynen, K. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McDannold, N. J.
Right arrow Articles by Hynynen, K. H.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE